Skip to content

LetJevDecide

Yes/no decision site that asks Jev and a fair in-browser coin and shows whether they agree. One relay call returns a verdict, the probability behind it and how significant a decision Jev thinks it is, behind crisis and not-yes-or-no gates. Plain HTML, no build step, no sign-up.

Stars
1
Language
JavaScript
License
MIT
Published by
Community
Author
choxos
Last verified
Oct 2026

About this project

Stuck on a yes or no decision? Type it in, then let Jev decide, or flip a fair coin. Use it at choxos.github.io/LetJevDecide or letjevdecide.xera.ac. No sign up, no install. Ask Jev. Jev is TypeSafe's System One model. It does not write text; it answers typed questions with probabilities. Your question gets a yes or a no, the probability of yes behind it, and how big a decision Jev thinks it is. Add details with the slider button in the question box and Jev weighs them too.

Quoted from the project's README. Read the full README.

Highlights

  • One POST to /v1/systemone asks four questions about the same state in parallel: crisis, is_yes_no, decide and importance, then thresholds them.
  • The coin is a crypto.getRandomValues flip in the browser, so nothing leaves the machine, and also-flip runs both deciders to compare them.
  • The relay is deliberately narrow: question up to 200 characters, details up to 1,000, origin allowlist, rate and token caps, no storage.
  • Plain HTML, CSS and JavaScript on GitHub Pages with no build step and no dependencies; history is localStorage, with no cookies or trackers.

Watch out

This is a toy and the README says so: for fun, not professional advice. The verdict rules are hand-set constants in relay/decide.js, with yes at 0.5 or more, anything within 0.1 of 0.5 flagged as on the fence, a crisis score of 0.5 or more suppressing the verdict in favor of pointing to help, and is_yes_no below 0.3 asking for a proper yes/no question; the model is pinned at jev-1.13.0. Only the coin is random, since Jev's answer depends entirely on the question and details you type, capped at 200 and 1,000 characters. The underlying coin-toss study is described by its own author as merely suggestive, with self-reported happiness and incomplete follow-up. Hosted use depends on the author's relay, so anyone uncomfortable with that should self-host: Node 20+, a TypeSafe key and an HTTPS reverse proxy with an explicit origin allowlist. Copy reviewed against the repository README on 2026-10-04, not from running it.

Questions about this entry

Short answers drawn from the linked project or page.

More Jev resources like this

23GitHub stars
Live race rig that labels 1,000 real app reviews with Jev typed questions against Gemini JSON output on identical batches, then replays recordings and re-ranks rows locally without extra model calls.
Jev Practices & Patterns#community#jevbest#catalog-import
Minecraft setup where Hermes plans at milestones, Jev picks one bounded harness action per step, and Mineflayer code owns validity and execution with no screenshots.
Jev Practices & Patterns#community#awesomejev#catalog-import
6GitHub stars
Watch Jev play chess against OpenRouter LLMs or Stockfish, or play Jev yourself, with every weighed move, likelihood, clock, and finished game saved for replay.
Jev Practices & Patterns#community#awesomejev#catalog-import
5GitHub stars
Inline verifier that checks LLM responses and agent plans with one Jev call across alignment heads and returns block, flag, or pass for gates and hooks.
Jev Practices & Patterns#community#awesomejev#catalog-import
6GitHub stars
Self-hosted Feishu bot plus web workbench where a text model proposes 2-5 candidate answers and TypeSafe Jev picks among them with probabilities.
Jev Practices & Patterns#community#awesomejev#catalog-import
4GitHub stars
OpenAI-compatible proxy that holds streamed tokens in a sliding buffer and scores the full criteria matrix before release, cutting the stream before a violating token reaches the client.
Jev Practices & Patterns#community#awesomejev#catalog-import

From the Jev guides

Original write-ups on what this entry is for.

Most Jev accuracy problems are state problems. What belongs in a state, what is a pre-made judgment to leave out, and why naming the slice you ask about matters more than the wording.
Jev Practices & Patterns#state#design#primitives
Read article
A Noul returns one number and no separate confidence. How to read the value, pin its boundary with criteria, split compound questions, and set a per-action threshold that matches the cost of being wrong.
Jev Practices & Patterns#noul#primitives#confidence
Read article
A Score is a weighted position, not a ruler. Why the number can land between levels, why levels are judged separately, and how to combine several atomic scores in code instead of one crowded rubric.
Jev Practices & Patterns#score#primitives#scoring
Read article
Back to all resources

From the Jev community

Posts from builders shipping with Jev right now.

Follow @typesafeai